基于字典中的值过滤Pythons Pandas数据帧

Question

我将一些数据作为pandas数据帧读入Python：

             Unnamed: 0  Initial_guess  Lower_bound  Upper_bound Estimated_or_Fixed  
      0          Ka              5     0.000001        10000          Estimated   
      2          Kd              5     0.000001        10000          Estimated   
      3          Ki              5     0.000001        10000          Estimated   
      5          Kr              5     0.000001        10000          Estimated   
      6        R1_I              5     0.000001        10000          Estimated   
      7         PR1              5     0.000001        10000          Estimated   
      8         PR2              5     0.000001        10000          Estimated   
      9       alpha              5     0.000001        10000          Estimated   
      10        Kcd              5     0.000001        10000          Estimated   
      12       Klid              5     0.000001        10000          Estimated   
      18    LR1R2_I              5     1.000000        10000          Estimated   

        Variable_type  
0   Kinetic parameter  
2   Kinetic parameter  
3   Kinetic parameter  
5   Kinetic parameter  
6   Kinetic parameter  
7   Kinetic parameter  
8   Kinetic parameter  
9   Kinetic parameter  
10  Kinetic parameter  
12  Kinetic parameter  
18         Species IC  

第一列unnamed: 0是参数。我有很多模型，每个模型都包含这些参数的不同组合。我的任务是通过删除模型中不存在参数的任何行来为每个模型筛选此表。我有每个模型的字典，它们包含参数。参数可以是两种类型，species IC或kinetic parameter。以下是第一种型号的字典示例：

^{pr2}$

我的代码：

def write_parameter_bounds_file(self):
    model1=self.all_models_dirs[0] #get first model from a list of model. I'll do it on the first model then generalize to the rest. 
    species=self.get_model_species(model1+'.xml') #get the species dct from this model
    parameters=self.get_model_parameters(model1+'.xml')#get parameter dct from this model
    param_info=self.read_parameter_bounds_template() #get all parameters from template. This is the pandas dataframe at the top. 
    estimated_species=[]
    estimated_params=[]
    for i in species.keys():
        print '\n'
        for j in param_info[param_info.columns[0]]:
            if i==j:
                estimated_species.append(i)
    for i in parameters.keys():
        print '\n'
        for j in param_info[param_info.columns[0]]:
            if i==j:
                estimated_params.append(i)
    param_list=estimated_params+estimated_species #This is a list of the parameters that need to be included in the output df

有人知道我如何使用param_list过滤原始熊猫的df吗？在

谢谢

Answer 1

您可以将函数isin与从字典生成的列表一起使用：

list_Species_IC = Species_IC.keys()

并获得dataframedf的子集。您可以通过函数reset_index重置索引。在

类似的方法可以用于dictionary Kinetic_parameter。在

^{pr2}$

总而言之：

Species_IC = {'R1': '2.7109e+02', 'R2': '1.2709e+02', 'R1_I': '2.7109e+03', 'R2_I': '1.2709e+03', 'LR1R2': '1.6913e+00', 'LR1R2_I': '1.6913e+01'}
Kinetic_parameter = {'Ka': '1.0000e+00', 'TGFb': '1.0000e-01', 'Synth': '1.0000e+00', 'PR1': '8.0000e+00', 'Sink': '0.0000e+00', 'PR2': '4.0000e+00', 'alpha': '1.0000e+00'}

list_Species_IC = Species_IC.keys()
list_Kinetic_parameter = Kinetic_parameter.keys()
list_IC = list_Species_IC + list_Kinetic_parameter
print list_IC
#['R1', 'R2', 'R1_I', 'R2_I', 'LR1R2', 'LR1R2_I', 'Ka', 'TGFb', 'Synth', 'PR1', 'Sink', 'PR2', 'alpha']
out = df[df['Unnamed: 0'].isin(list_IC)].reset_index()
print out
#   index Unnamed: 0  Initial_guess  Lower_bound  Upper_bound  \
#0      0         Ka              5     0.000001        10000   
#1      4       R1_I              5     0.000001        10000   
#2      5        PR1              5     0.000001        10000   
#3      6        PR2              5     0.000001        10000   
#4      7      alpha              5     0.000001        10000   
#5     10    LR1R2_I              5     1.000000        10000   
#
#  Estimated_or_Fixed  
#0          Estimated  
#1          Estimated  
#2          Estimated  
#3          Estimated  
#4          Estimated  
#5          Estimated